Getting to the “Why” in Global Development

The CUBES Framework and Toolkit

Development experts, doctors, governments, and philanthropists like to tout the great strides that have been made in global health and development in recent decades. The Green Revolution used science to double yields of cereal crops between 1960 and the mid-1980s, helping avoid widespread famine as the world’s population mushroomed. Polio has been eradicated in all but three countries, thanks to a relentless campaign of vaccination. Humans have plenty about which we can pat ourselves on the back.

Yet there’s a puzzling aspect to this story of progress. Technical solutions exist for many of the global health issues facing humanity, yet problems persist. Even though condoms are excellent at preventing sexual transmission of HIV, 1.8 million adults around the world became infected with the virus in 2017, the vast majority through unprotected sex. Giving birth at a medical facility rather than at home is clearly safer for a woman and her newborn, yet in many developing countries home births continue to be the norm, and newborn mortality rates are stubbornly high.

These examples share a key characteristic: positive health outcomes are dependent upon an individual’s behavior. However, behavior is not simply the product of personal decisions, but is shaped by an intricate web formed of other people, systems, and the wider culture. This makes the question of why people behave as they do fundamental to understanding how to change their behavior.

Behavior change has long been a key goal of many health and development programs, and it is only growing more urgent as donors and governments try to squeeze the greatest possible impact from their investments. If behavior change is hard for individuals (just think of those get-fit resolutions we fail to keep year after year), it’s even more challenging for programs that are trying to influence the behavior of thousands or millions of people.  

First, program designers trying to promote a positive behavior must investigate whether people are not yet aware of it, or are aware but lack the intention to do it, or have the intention but fail to follow through. Next, they must elucidate the many drivers – both external and internal – that influence whether people progress through those stages to finally take action. Each driver potentially requires a separate intervention to change people’s behaviors.

A plethora of models has sprung up in recent decades to help understand this complex and dynamic ecosystem of people, places and information. There is a bewildering diversity of approaches and techniques for conceptualizing and analyzing behavior. Today, program designers tackling a global health issue can draw on psychology, sociology, behavioral economics, or market research for insights that may prove complementary – or just conflicting.

We faced these challenges in our work on voluntary male circumcision for HIV prevention in southern Africa and saw this to be a consistent gap across development programs. We founded Surgo to focus on the why of human behavior in order to change minds, systems, and outcomes in global development. Previous work on generating demand for voluntary male circumcision had articulated how the stages of change combine with external and internal drivers to enable or hinder behaviors. In our work at Surgo over the past few years, we’ve developed this into a comprehensive framework for understanding behavior change. It draws from our on-the-ground experience, along with other models, literature reviews, and consultations with a range of experts. In designing our framework, we favored approaches that have been used most extensively across different fields, and that have the best evidence for predicting behavior.

But simply identifying the relevant set of behavioral drivers isn’t enough: you have to be able to know which ones your program should target to have the greatest effect. So we’ve created a practical toolkit of methods to measure the significance of these drivers. We’ve named the resulting framework and toolkit with an acronym, CUBES: to Change behavior, Understand Barriers, Enablers and Stages of change.

Our framework recognizes that behavior emerges from a complex set of interactions between people and their environments. It provides a kind of checklist that helps program designers make sure they’ve considered all the likely elements of an issue. The external environment contains plenty of behavioral drivers. Policies, laws, and rules may enable or hamper behaviors. Infrastructure plays a part: the decision to give birth at a clinic won’t count for much if the facility is too far away, or the roads are too poor to get there. And antimalarial bed-nets to protect sleepers from mosquito bites won’t help if they’re not easily available or affordable.

The CUBES Framework


But just as important are internal drivers of behavior: what do people think about whether bed-nets are easy to use, whether they really work, and how risky it is to sleep without one? Alongside these conscious perceptions, psychology has shown that our behaviors are often influenced by unconscious rules of thumb or shortcuts that bias our decision-making. Our emotions play an important role, too: for example, a program in Ghana to encourage the use of soap found that educating people about its health benefits had little impact, but messages emphasizing feelings of disgust around dirty hands produced the desired change in behavior.

Varied personality types must also be taken into account: someone who is conscientious may respond to a different approach than a person whose predominant trait is extraversion. And since people exist within social networks, the influences of family, friends, the local community, and wider society also have a part in enabling or constraining behaviors. Traditional and social media often play an important role here.

There are many potential drivers, but which should be targeted to change behavior effectively and efficiently? Many development programs try to figure this out by defaulting to interviews, focus groups and quantitative surveys. The CUBES toolkit encourages programs to think again by offering a compendium of methods that go beyond the most commonly used approaches. Existing methods can be improved upon: quantitative surveys can be checked against CUBES to make sure they will capture data on all the barriers and enablers of behavior change.

Second, information gaps can be filled using less-familiar methods. For example, journey mapping is a technique from market research that zeroes in on the stages of change. “In vitro” experiments can test specific drivers and narrow down potential interventions, so that only the most likely candidates are tested in the field. For programs with the resources to use machine learning (artificial intelligence), and valid data to feed into it, computer simulations can explore multiple “what-if” scenarios, teasing out the importance of individual variables and showing the likely effect of different interventions.

After gathering data using a range of methods, researchers can use the CUBES framework to structure their findings and identify the best content and potential targets of an intervention. In a complex real-world setting, it’s inevitable that not all behavior change interventions will be successful. But knowing which drivers of behavior are important in a given situation improves the chances of choosing suitable intervention types. It reduces the degree of trial and error.

Surgo used the CUBES framework to examine how decisions around family planning are made in Uttar Pradesh. We looked at existing data in a new way and discovered a finding that others had missed: households were already well aware of effective family planning methods, but they indicated a low intention of using them. There were strong community norms to demonstrate fertility by having many children, and community health workers and nurses tended not to discuss family planning. Family members perceived the risk of unintended pregnancies to be low; they felt shame around obtaining family planning goods, and believed strongly in the possibility of adverse side-effects.

Interpreting the results using CUBES, it was clear that interventions to increase awareness of family planning were unnecessary. Instead, the barriers that needed to be addressed were household members’ fears of adverse side-effects, and social norms about childbearing. We also found that community health workers and nurses were a weaker influence than assumed, which meant looking at other channels to influence households. Husbands, for example, played a larger role in family planning decisions than expected. We are now conducting large-scale surveys to measure the relative importance of these behavioral drivers.  

The CUBES framework and toolkit is an important step toward making development programs more effective and efficient. It helps program designers in global development and beyond avoid the trap of grasping at immediate solutions that may not address the underlying issue, and lowers the risk of wasting time and resources. In short, our framework helps examine and understand the complete ecosystem underlying human behavior, and offers an expanded approach to gathering data to really understand why people behave as they do.